WO1986005203A1

WO1986005203A1 - Dna encoding anion transport protein

Info

Publication number: WO1986005203A1
Application number: PCT/US1986/000450
Authority: WO
Inventors: Seth L. Alper; Ron R. Kopito
Original assignee: Whitehead Institute For Biomedical Research
Priority date: 1985-03-07
Filing date: 1986-03-06
Publication date: 1986-09-12
Also published as: EP0216853A1

Abstract

The cloning and sequencing of cDNA encoding anion transport proteins. In particular the cloning and sequencing of cDNA encoding erythroid murine Band 3, which can serve as the basis for development of probes to be used in detecting the presence or absence of genetic material encoding protein which transports chloride ions and bicarbonate ions, are disclosed. The use of cDNA encoding murine Band 3 in the detection and isolation of Band 3-like proteins and DNA encoding those proteins in other tissues is also described.

Description

DNA ENCODING ANION TRANSPORT PROTEIN

Description

Technical Field This invention is in the field of molecular biology and in particular relates to anion exchange across cell membranes.

Background Art

The composition of the fiuid inside of cells differs from that outside of cells. A lipid mem¬ brane separates these tv/o compartments and prevents their compositions from equilibrating. However, this membrane .is not completely impermeable, but, rather, selectively permeable. Specialized pro- teins, located in the membrane, confer limited permeability to specific salt ions or nutrients. The pattern of specificity varies according to the functions and requirements of particular cell types. Chloride ions (Cl~) are transported across the plasma membrane by themselves (uniport transfer) or with another ion. For example, they may move in the

4- + same direction as Na K (symport) or in the opposite direction across the erythrocyte plasma membrane as bicarbonate (HC0,~) (antiport) . Little is known about Cl transport across cell membranes, especi¬ ally in epithelia; even less is known about the molecule(s) which mediate either the active or passive exchange of chloride in these tissues.

It has recently been suggested that a defect in Cl transport is responsible for the disturbance of fluid and electrolyte transport which is thought to be central to the etiology of cystic fibrosis (CF) . Quinton, P.M. In: Fluid and Electrolyte Abnormali¬ ties in Exocrine Glands in Cystic Fibrosis. (P.M. Quinton et al., ed.)- pp 53-77. (1982). The physio¬ logical defect seems to be an inability of Cl -to cross the specialized epithelial cells that are components of exocrine tissues. Quinton has pro¬ vided evidence that abnormally low Cl^" permeability in CF leads to poor reabsorption of sodium chloride (NaCl) in the sweat duct and this produces the high concentration of NaCl in the sweat, which is charac¬ teristic of CF patients and which serves as the basis for diagnostic tests. Quinton, P.M. Nature, 301; 421-422. (1983). Knowles and coworkers have also demonstrated a defect in luminal Cl perme- ability in nasal epithelium from CF patients.

Knowles, M. et al. Science, 221; 1067-1070. (1983).

In both cases, the defect was shown to be the inability of Cl to cross the epithelial cells of an organ known to be affected in CF. These findings have led to speculation that the genetic defect underlying CF involves an as yet unidentified protein which transports Cl^" across plas a membranes. The protein identities of many cation ion transport activities have been known for a long time. However, only one anion transport protein of the plasma membrane has been identified;

5 this protein is Band 3.

Disclosure of the Invention

This invention is based upon applicants' cloning and^' sequencing of cDNA encoding murine Band 3. Band 3 is the major glycoprotein of mammalian

10 erythrocytes and functions as an anion antiport in that it mediates the one-for-one exchange of chloride (Cl ) and bicarbonate (HCO, ) across the erythrocyte plasma membrane. Applicants have used RNA from erythroid precursor cells obtained from

15 mouse anemic spleen to construct a cDNA library.

Through screening of the clones using a polyclonal antibody against purified mouse erythrocyte Band 3, they have identified clones containing DNA inserts encoding segments of the murine Band 3 protein.

20 These DNA inserts were isolated and sequenced; the longest of these, which contained an 1800bp cDNA insert, was used as a hybridization probe to screen a second cDNA library. The clone in this second cDNA library which contained the longest cDNA insert

25 (pB399) was identified and isolated from this second #^* library. The library was rescreened using pB399 and a clone containing the entire 2900bp open reading frame encoding mouse Band 3 was identified. The nucleotide sequence of each of the cDNA clones was

30 determined. The sequence of 4257 nucleotides encoding Band 3 was compiled from the three clones and is similar in size to that of the mRNA obtained from the erythroid precursor cells. - d -

Development of probes useful in detecting the presence in cells of genetic material encoding proteins which transport anions across cell mem¬ branes can be based on this information. In particular, cloning of cDNA encoding murine Band 3 can serve as the basis for making probes to be used in detecting the presence in cells of mRNA encoding proteins which transport chloride, bicarbonate and other anions. Such probes have been used to detect and isolate proteins similar to erythroid Band 3, as well as the cDNAs encoding the proteins, in mouse kidney cells and rat kidney.

Proteins encoded by the cDNA have been produced using recombinant DNA methods. For example, the complete Band 3 polypeptide has been expressed by clone pB3SP4, which has the entire 2900 bp open reading frame encoding mouse Band 3 protein.

Applicants* cloning of cDNA encoding murine Band 3 protein in recombinant vector host systems also provides the basis for determination of the nucleotide sequence of the cDNA.

Brief Description of Drawings

Figure 1 is a schematic representation of cDNA encoding murine Band 3 glycoprotein with its restriction enzyme sites and of cDNA inserts encoding a segment of the murine Band 3 glycoprotein in representations of clones pB33, pB399 and pB3SP4, respectively. _^•**. -

Figure 2 is a representation of the nucleotide sequence of Band 3 cDNA and the deduced amino acid sequence of the protein encoded by it.

Figure 3 presents a comparison of the alignment 5 of amino acid sequences in fragments of human Band 3 glycoprotein with the deduced amino acid sequence of portions of the murine Band 3 protein of Figure 2. Figure^' 4 is a schematic representation of the three structural domains of murine Band 3 protein. 1^*0 Figure 5 is a schematic representation of a model for orientation of Band 3 protein in the plasma membrane.

Figure 6 is a block diagram of one embodiment - of an ssay for genetic material encoding anion 15 transport protein which could be produced according to the methods described herein.

Figure 7 is a representation of the nucleotide sequence of cDNA isolated from murine kidney cells which encodes a Band 3-like protein. 20 Figure 8 is a representation of the nucleotide sequence of cDNA isolated* from murine kidney cells which encodes a Band 3-like protein (line 1 and subsequent odd numbered lines) and the nucleotide sequence of Band 3 cDNA encoding its carboxy- 25 terminal 29' amino acids (line 2 and subsequent even ^"" numbered lines) .

Figure 9 is a^' representation of the deduced amino acid sequence of the protein encoded by the nucleotide sequence represented in Figure 7 (line 1 30 and subsequent odd numbered lines) and of the deduced amino acid sequence of murine erythroid Band 3 protein (line 2 and subsequent even numbered lines) . Best Mode of Carrvinσ Out the Invention Band 3 Glycoprotein

As mentioned, Band 3 is the major membrane glycoprotein of the mammalian erythrocyte. In addition to its role as an anion antiport. Band 3 serves another essential function in the erythro¬ cyte. It binds tightly to ankyrin and anchors the mesh-like network (composed of the proteins spectrin and actin) which constitutes the erythrocyte sub- membrane cytoskeleton. Band 3 has been^' shown to have two distinct structural domains into which its anion exchange activity and cytoskeletal inter¬ actions are segregated.

Band.3 has 929 amino acids. The a inό-terminal (approximately) 400 are hydrophilic and face the cytoplasm. This negatively charged domain possesses the high-affinity binding site for ankyrin as well as, in human Band 3, binding sites for hemoglobin and. several glycolytic enzymes. Removal of the cytoplasmic domain of Band 3 by proteolytic cleavage of erythrocyte ghosts results in no loss of anion transport activity. This demonstrates that the transport activity resides entirely within the C-terminal (approximately) 500 amino acids of the protein.

Studies suggest that the transport domain traverses the membrane at least nine times. It has not been possible to determine the accuracy of this model, however, because only fragmented amino acid sequencing data is available for this region of the human Band 3 protein. Braell and Lodish have demonstrated that human Band 3 and mouse Band 3 proteins are very similar in size and biochemical properties. Braell, VI . A. and Lodish, H. F. Journal of Biological Chemistry, 256: 11334-11337. (1981).

Isolation of Clones Having cDNA Encoding Band 3

5 Protein

In mice, anemic stress induces proliferation in the spleen of erythroid precursor cells that actively synthesize erythrocyte proteins. Poly- adenylated RNA was obtained from these cells. Using

10 a "shotgun" approach, the RNA was used to construct a cDNA library* in the expression vector, lambda- gtll. Young, R.A. and Davis, R.V7. , Proceedings of the National Academy of ^'Sciences, U.S.A., 80; 1194-1198. (1983). The cDNA library was screened

15 with a polyclonal antibody against purified mouse erythrocyte Band 3 which can immunoprecipitate Band 3 from an _in vitro translation reaction programmed with total RNA from anemic mouse spleen. Initial screening yielded several positive clones, the DNA

20 inserts from which hybridized to each other and to a 4300 nucleotide RNA in total spleen RNA. pB33, the longest of these, contained an 1800 bp cDNA insert. This insert was subsequently sequenced and shown to encompass only about 280 bp of the C-terminal "J^* 25 protein coding region; it extends throughout the entire 3' untranslated region of about 1500 bp. (Figure 1) pB33 was used as a hybridization probe to screen a second cDNA library which was con¬ structed to optimize the yield of full-length

30 transcript. (See Example 1) Clone pB399 contained the longest insert (2.7kb). Nucleotide sequencing revealed that it did not contain the N-terminus of native Band 3 (as assessed by the homology of the deduced amino acid with that of the N-terminus of human erythrocyte Band 3) . The library was re- screened with a 5' fragment of pB399; this resulted in the isolation of clone pB3SP4, whose nucleotide sequence was also determined. It contains 127 bp of 5' untranslated region and the entire 2900 bp open reading frame encoding mouse Band 3 (Figure 1) . Clone pB3SP5, which has the same nucleotide sequence as that of pB3SP4, inserted in the reverse direc¬ tion, has been deposited with the American Type Culture Collection (Rockville, MD) under deposit number 53047.

Determination of the Nucleotide Sequence of cDNA

Encoding Band 3 Protein The nucleotide sequence of the isolated cDNA encoding murine Band 3 protein was determined by the DNA sequencing procedures of Sanger and of Maxam and Gilbert. F. Sanger et al. Proceedinσs of the National Academv of Sciences,

U.S.A., 74; 5463-5467. (1977) Maxam, A. and Gilbert, . Methods in Enzymology, 64; 499-560 (1980). See Example 1.

Seσuence Ho oloσv of Mouse Band 3 and Human Band 3 Figure 2 illustrates the sequence of 4257 nucleotides compiled from the cDNA clones pB33, pB399 and pB3SP4. The size of the full-length cDNA represented by the three clones is similar to that of the RNA. The overlapping regions of the three clones have identical sequences. A single open

reading frame extends for 929 codons from the ATG at nucleotide 1 to TGA at base 2919. Two in-frame ATG codons occur at the beginning of the open reading frame (residues 1 and 4) . The first Met (methi- onine) codon was designated the initiator because it is the first in-frame ATG downstream of the (in phase) stop codon at base -63. The sequences flanking this ATG, but not the other, are homologous to the highly conserved sequence that flanks func- tional initiation sites in eukaryotiσ mRNAs; AXX AUG G. A polypeptide of about 103,000 dalton molecular weight is predicted on the basis of the cDNA; this is in good agreement wi^h estimates of 90,000 to 100,000 obtained from SDS-polyacrylamide electrophoresis of the native human and mouse proteins. The predicted C-terminal amino acid of the mouse sequence is valine, which is also the C-terminal residue of the human protein. Eleven of the C-terminal 12 resides predicted from the mouse cDNA sequence are identical with those determined by carboxypeptidase-Y digestion of human Band 3. Parts of the amino acid sequence deduced from the cDNA sequence are closely homologous to that of the several known fragments of human erythrocyte Band 3. (Figure 3) This is clear evidence that the isolated clones contain cDNA that encodes the mouse erythro¬ cyte anion exchange protein.

The most striking ho ology occurs between residues 456 and 492 of mouse Band 3 and the 38 amino acid fragment from the human protein designated H3 in Figure 3. There is only a single conservative (leu - Val) Stbstitution between the two sequences; there are no insertions or deletions. Excellent alignment is also observed between the deduced sequence of mouse Eand 3 and the sequence of a peptic fragment (P5) of the human protein desig- nated H4 (Figure 3) ; only 5 of the 72 residues are different and alignment of the sequences required no insertions or deletions. The sequence of the amino-terminal 201 residues of human Band 3, desig¬ nated Hi, also exhibits a high degree of homology with the sequence deduced from the mouse Band 3 clone. Although substantial, this homology is lower overall than either the two C-terminal sequences discussed above. Computer-assisted alignment, using the mutation data matrix, which emphasizes evolution- ary relationships in scoring amino acid homologies, suggested that mouse Band 3 is slightly longer than its human homologue. The homology is lowest at the extreme N-terminus of the protein, and seems to occur in clusters distributed throughout the seg- ment. (Figure 3)

Mouse Band 3 Structural Domains and Membrane Orienta¬ tion

Mouse Band 3 can be roughly divided into 3 domains. Kyte, J. and Doolittle, R.F., Journal of Molecular Biology, 157: 105-132. (1982). The

N-terminal 420 residues constitute the first domain, the hydrophilic cytoplasmic domain; it has a nega¬ tive net charge because of its 65 acidic and 37 basic residues. The central 450 amino acids conεti- tute the second domain and are predominantly amphi- -ι ι -

pathic. They are intimately associated with plasma membrane lipid. In this domain, groups of hydro- phobic residues are interspersed with polar, pre¬ dominantly basic, residues. These are shown sche ati- cally in Figure 4. The third domain of Band 3 encompasses the extreme 32 C-terminal residues. Eleven of these are either glutamate or aspartate, making it unlikely that the C-terminus is buried within the lipid bilayer. The cytoplasmic domain. The binding of hemoglobin and glycolytic enzymes to the cytoplasmic domain of human Band 3 is mediated primarily by the first 11 residues of the Band 3 polypeptide. The fact that mouse erythrocyte Band 3 does not bind glyceralάehyde- 3-phosphate dehydrogenase is^* consistent with the total lack of homology between mouse and human sequences in this N-terminal undecapeptide.

Other features of the cytoplasmic domain of Band 3 are more conserved. Human erythrocyte Band 3 possesses 6 cysteine residues, 5 of which are reactive with N-ethyl-maleimide. Mouse Band 3, as predicted from the cDNA sequence, also has 6 cysteines. (Figure 4) Three of these are located within the cytoplasmic domain (CH65) , which is consistent with the number determined in studies on human Band 3.

Based upon studies on pH-dependent reversible conformational equilibrium, Low et a_l. proposed a domain structure for the cytoplasmic domain of Band 3. They suggest that the cytoplasmic domain is a homodi er with a "regulated hinge" rich in proline resides somewhere in the middle of CH65. Low, P.Ξ. et al. , Journal of Biological Chemistry, 259: 13070-13076. (1984) These proline residues ap¬ parently disrupt the folding of the polypeptide in this region, rendering it susceptible to prote- olysis. It is noteworthy that these proline resi¬ dues (at positions 161, 188, 201 and 204) are all conserved between human Band 3 and mouse Band 3. (Figure 3) Figure 4 shows them to be clustered around a known trypsin cleavage site at position 194. Two other proline-rich regions in the predicted mouse sequence are found at the known primary sites of intracellular and extracellular proteolysis, between residues 361-368 and 567-586, respectively. (Figure 4) The purported antigenic region of Band 3 occurs between the proposed "hinge" and a region rich in tryptophan, (residues 89-119) ; this is also highly conserved. Monospecific polyclonal anti¬ bodies against total human Band 3 recognize deter¬ minants located within a 20 kD fragment of the cytoplasmic domain; this site is sufficiently distant from the N-terminus so that the binding of antibody does not interfere with the binding of hemoglobin. The region between residues 125-138 is only 57% homologous between human and mouse; the residues flanking it on either side are 90-95% homologous. This observation is consistent with the lack of cross-reactivity of anti-human Band 3 antisera with the mouse protein and vice-versa.

The ankyrin binding site on Band 3 has not been identified. Because antibodies against Band 3 compete with labelled ankyrin for binding, Low e_t a_l tentatively place it between the "hinge" region and the "IgG" region. Low, P.S. et al. , Journal of Biological Chemistry, 259; 13070-13076 (1984) . It is likely that the high-affinity ankyrin-binding site is highly conserved among different species, and perhaps, among different non-erythroid ankyrin- binding proteins. Figure 4 shows the suggested position of this binding site.

The membrane-associated domain. Figure 4 indicates that the C-terminal segment contains ten long hydrophobic stretches, labelled A-J. Regions with hydropathy averages exceeding 1.5 are usually tightly associated with lipid, and, in the case of several proteins, span the membrane in an alpha- helical configuration. The two-dimensional model for Band 3 represented in Figure 5 indicates that seven of these peaks correspond to sequences of amino acids that span the phospholipid bilaver once (A, C, D, E, F, G, I) and that two (B, I) and possibly a third, (J) , span twice. Peaks A, C, E, G, and II, designated single- spanning regions 1, 4, 6, 8 and 9, respectively, correspond to continuous stretches of 20-24 ex¬ tremely hydrophobic amino acid residues. These readily fit the criteria for forming membrane- traversing alpha-helices. These 5 helices^*contain few proline residues, which would disturb the regular alpha-helical structure, and fev.' amino acids with either charged or polar side-chains.

Digestion of intact human and mouse erythro- cytes with chymotrypsin cleaves Band 3 into a carboxyl-terminal glycosylated peptide, CII35, and an N-terminal non-σlvcosvlated CH65. (Ficrure 4) Se- quence data on the N-terminus of human CH35 posi¬ tions this site between the putative membrane- spanning segments 5 and 6. Therefore, this segment must be extracytoplasmic. As deduced by Brock e_t aJL. from their sequence, the region between segments 7 and 8 is extracellular, and the residues between 6 and 7 are exposed to the cytoplasm. Of the two potential sites (Asn-X-ser/thr) for attachment of the single N-linked oligosaceharide of Band 3, the one at residue ^*611 corresponds to the site in the human protein which has been shown by Brock e_t al. not to contain carbohydrate. Brock, C.J. et al. , Biochemical Journal, 213; 577-586. (1983) There¬ fore, the site at Asn 660 must be -the attachment site and thus extracellular. This analysis, taken together with the finding that the human Tyr 646 is a substrate for extracellular radioiodination, supports the conclusion that the region between helices 7 and 8 is extracellular. The region encompassing segments 1 to 5 cor¬ responds to part of the 17-22 kD chymotryptic fragment, CH17, that is generated by digestion of ghosts. (Figure 5) Because it results from both an intra- and extracellular proteolytic cleavage, it must span the membrane an odd number of times. Studies with surface-specific radiolabelling re¬ agents demonstrate that this fragment traverses the bilaver a minimum of three times. Jennings, M.L. and Nicknish, J.S., Biochemistry, 23: 6432-6436. (1984). Three lysines in human CH17 can be reductively methylated in intact cells by a membrane impermeant reagent and thus must be exofacial. Two of these are within the C-terminal CNBr fragment of CHI7. There are only three lysine residues in murine CH17 (at positions 449, 558 and 561). In light of the high degree of homology between the human and mouse proteins in^' at least part of CH17 (H3) , it is believed that the positions of these lysine residues in human and mouse CH17 are the same.

One of the two lysines in the C-terminal CNBr fragment of human CH17 (CNll) is the one which binds extracellular 4,4'-diisthiocyano2,2'-dihydrostilbene disulfonate (H-DIDS) , a bifunctional reagent which is a potent inhibitor of aniόn transport and cova- lent crosslinker of CH17 and CH35. Segment #5 forms the C-terminus of mouse CH17. If it is assumed that there are 3.6 residues per turn, use of a two dimensional "wheel diagram" , indicates that segment £5 could form a membrane-spanning structure in which both Lys 558 and 561 of mouse CH17 project near the extracytoplasmic surface of the membrane from the same side of an amphipathic alpha-helix as the two acidic residues, Glu 554 and Asp 565. The remaining lysine in mouse CH17 is located near the N-terminus of the fragment (Lys 449) . This residue is exofacial in human erythrocytes and it is likely that Lys 449 in the mouse sequence must also face the extracytoplasmic surface. This establishes that region A forms membrane-spanning domain #1 shown in Fiσures 4 and 5. If it is assumed that peaks A, C, and D of the hydrophobicity plot correspond to three membrane- spanning segments, 1, 4, and 5, then either region B (residues 453 to 491) either does not span the membrane, or does so twice. Hydrophobicity peak B is quite broad, it encompasses 38 moderately apolar residues. These residues can be arranged in a structure composed of two anti-parallel amphipathic helixes in which Pro 477 forms a turn within the membrane or just at the interface between the bilayer and the cytoplasm. Analysis of the helical amphiphilicity suggests that both of .the helices formed by this region of the protein would have all of their polar residues projecting from a single face of each helix. Thus, it appears that CHI7 has 5 membrane-spanning alpha-helices, three of which three (2, 3 and 5) are amphipathic.

The human counterpart to murine membrane- spanning segments 6 and 7 is a pepsin fragment (P5) , which has been sequenced by Brock et al . They concluded that #7 forms an amphipathic helix through the lipid bilayer. The importance of this region in the anion exchange process is supported by the observation that papain cleavage of intact erythro- cytes inhibits anion exchange and by the substantial sequence homology between the mouse and human proteins in this region. (Figure 3)

Virtually nothing is known about the trans- membrane orientation of the C-terminus of Band 3 corresponding to hydrophobic peaks G-J. The model represented in Figure 5 proposes 4 or 5 membrane crossings. The region corresponding to the 43 residues of peak I of a hydro in Figure 4 resembles membrane spans 2 and 3 in its ability to form an alpha-helical hairpin in which Arg 800 is situated just at the cytoplasmic surface and the C-terminal 24 amino acids return to the extracellular surface. The proline residues at positions 796 and 802 may be involved in forming the "hairpin" structure. These two spanning segments, along with #12, could form moderately amphipathic helices. Following an extremely polar stretch between residues 832 and

850, there is a short region of 18 mixed hydrophobic and polar residues which forms the N-terminal part ' of.peak J. (Figure 4) This is the longest stretch of apolar amino acids between position- 850 and the C-terminus of the protein. It is not clear whether this segment spans the membrane once or twice. If it spans the membrane once, as an amphipathic alpha-helix, this would position the C-terminus of the protein which would be positioned inside the cell. (#12, Figure 5) Alternatively, the protein may traverse the membrane a second time, leaving the C-terminus outside the cell. These two possibilities are shown as dotted lines in the model. It is known that the lysine residue in CH35 which is involved in the H^DIDS crosslinking of this fragment to CH65 is located somewhere within the C-terminal papain fragment, P25. In the mouse, P25 contains 11 lysines. Lysines 709, 713, 716 and 910 cannot bind extracellular H _.DIDS because they must be located at the cytoplasmic face of Band 3. It is possible that one of the 2 lysines between helices 7 and 8 (Lys 649 and 657) is involved in H₂DIDS crosslinking because of their proximity to the region of the protein where papain cleavage is associated with an alteration in anion transport. (Figure 5) It is also possible that other exofacial lysines closer to the C-terminus of P25 are involved in H₂DIDS binding.

The 3'^' untranslated region. There is a con¬ sensus polyadenylation signal located 18 bp upstream of the 3' terminal run of 8 A's in the Band 3 cDNA nucleotide sequence. (Figure 2) This suggests that clone pB33 spans the entire 3' untranslated region of the Band 3 mRNA. Application of Staden's posi¬ tional base preference analysis reveals that nucleo- ^' tides 1-2916 have a high statistical probability of encoding a polypeptide, but that the 1500 bp at the 3' end is consistent with that of a non-coding region. Staden, R. , Nucleic Acids Research, 12; 499-504. (1984) The function of such a long un- translated region at the 3 ' end of a message is unknown, but the mRNAs for other membrane proteins possess long (more than 1 kb) 3' untranslated regions: HMG-CoA reductage, the sodium channel from Electrophorus, and the receptors for LDL, EGF, and transferrin.

The 3' untranslated region of Band 3 mRNA has short tandem repeats, between positions 3490 and 3550. (Figure 2) This region is composed of 10 direct, perfect repeats of the tetranucleotide ATTC. Within this sequence are also 5 imperfect direct repeats of this sequence in which 1 out of 4 bases is different. A computer search revealed no signi- ficant homology between this sequence and the contents of the Dayhoff database. The functional significance of this region is unknown. The 3' untranslated region of the human LDL receptor mRNA possesses multiple Alu sequences which occur as 300 bp units oriented in a head-to-tail configuration.

Formation and Use of Probes Which Are cDNA Sequences

Which Encode Band 3 Protein

A. Probe Formation sequences which encode Band 3 protein can be intro¬ duced into a plasmid host which is a high copy for the vector. For example, cDNA can be inserted into pUC vectors (e.g., pUC13, pUC18,. pUCl9 or their derivatives) . The vectors are then used to trans- form cells, such as E.coli (e.g., E.coli JM103). The transformed cells are then cultured. The plasmid cDNA is then isolated; this can be done, for example, by centrifugation through a cesium chloride gradient, followed by ethanol precipitation. It will be recognized by those skilled in the art that other materials and methods than those specifically described herein can be used. For example, although bacterial plasmids have been described, other recombinant DNA vectors, such as phages and yeast vectors, could be used. In addi¬ tion, although Band 3 cDNA is described, probes, homologous to cDNA encoding other proteins which mediate ion transport across cell membranes can be made. Production of the probes can be carried out by several methods known to those skilled in the art. For example, the cDNA of interest can be cloned as described above or can be produced through enzymatic polymerization (e.g., through the use of a poly- merase such as Sp6 and a vector which has an Sp6 promoter) . Enzymatic polymerization can be used to produce single stranded (ss) DNA, double stranded (ds) DNA and RNA. The probes, which can be DNA (ss or ds) , RNA, or oligonucleotides, can be labelled by well-known methods. For example, they can be labelled through the addition of radioactive materials (e.g., 32P) or insertion of radioactive nucleotides. Alterna- tively, they can be labelled by chemical means.^' For example, biotin can be conjugated to the cDNA probe. The DNA sample can be attached to a filter and the biotin-labelled cDNA added. The addition of avidin (which binds to biotin) conjugated to the enzyme peroxidase, along with peroxide and an indicator will result in a color-producing reaction at those locations at which the biotin-labelled cDNA probe has hybridized to homologous sequences in the DNA sample. The cDNA probe which encodes murine Band 3 protein can be used to probe other tissues (e.g., kidney, spleen, heart, lung, stomach) for the detection of genetic material homologous to it and believed to encode proteins which transport Cl~ in epithelia. As described in Example 3, probes, which are fragments of mouse erythrocyte Band 3, have been used to detect and isolate Band 3-like protein and cDNA encoding it from mouse kidney cells and rat kidney. Specifically, two fragments of mouse erythroid Band 3 which encode, respectively, bases 3748-2368 and bases 2369-3132 have been used to probe a cDNA library constructed from murine renal cells for homologous sequences. cDNA was detected and isolated in this manner; the cDNA was sequenced and shown to encode 291 amino acids of the carboxy- terminal end of a protein exhibiting approximately 66% homology with the same portion of murine erythroid Band 3 protein. E. Coli strain GT-1 containing the plasmid pTALI, which is a cDNA of about 1150 bases of mouse erythrocyte Band 3 in¬ serted into the EcoRI site of poly linker of the plasmid pUC13, has been deposited with the American Type Culture Collection (Rockville, MD) under deposit number 67008. In addition, use of similar probes has demonstrated the presence in mouse and rat kidney cells of cDNA encoding mRNA whose trans¬ cripts have been shown by hybridization analysis to be even more similar to erythroid Band 3 mRNA.

Thus, it is possible to use probes which are cDNA fragments of murine Band 3 to detect the presence of mRNA encoding such proteins in other tissues, as well as to isolate the cDNA and the protein. The mouse cDNA probe can be used to screen, in addition to mouse tissue, human tissues (or other animal tissues) for the presence of homologous genetic material (e.g., cDNA or genomic DNA). The probes produced in this fashion, from mouse tissue, human cells or other mammalian cells, can be used to screen human genomic DNA for sequence polymorphism. The DNA or RNA probes can be used for testing samples (e.g., amniotic fluid, tissues) for the presence or absence of DNA or RNA encoding ion transport proteins. One specific way in which this can be done is as follows: the DNA from the tissue sample is isolated (by known methods) and cut by the use of restriction enzymes. The resulting fragments are separated (e.g., by gel electrophoresis) and blotted onto a filter. The- DNA or RNA probe is then applied to the filter and the presence or absence of homologous sequences is determined by the presence or absence of different bands on the gel. Differ¬ ences which occur v/ithin a band can be detected by differential hybridization. Production and Use of Antibodies to Detect Transport

Proteins or a Segment Thereof

In addition to the use of probes for the detection of presence or absence of genetic material encoding ion transport proteins, it is possible to produce antibodies to the transport protein. Such antibodies, which can be polyclonal or monoclonal, can be used to detect the presence of the transport protein, a related peptide or a segment thereof. Polyclonal antibodies can be produced by employing protein to immunize a host, such as a rabbit, and antibodies to the protein can be col¬ lected from serum obtained from the host. Mono¬ clonal antibodies can be produced employing cells which produce antibodies to the protein produced by the isolated gene in typical fusion techniques for forming hybridoma cells. Basically, these tech¬ niques involve the fusing of the antibody-producing cell with a cell having immortality, such as a myeloma cell, to provide a fused cell hybrid which has immortality and is capable of producing the desired antibody, in this case an antibody to the protein coded for by the isolated gene. The hybrid cells are then cultured under conditions conducive to the production of antibody which is subsequently collected from the cell culture medium. Such techniques for producing monoclonal antibodies have been well described in the literature. See, for example, U.S. Patent Nos. 4,172,124 an 4,196,265 issued to Hilary Koprowski et al., the teachings of which are hereby incorporated by reference. Other techniques for immortalizing antibody-producing cells for the purpose of producing monoclonal - antibodies can also be employed, such as by trans¬ forming such antibody-producing cells with viruses.

The antibodies so produced can be used in immunochemical assays. These include "sandwich" or "two-site" immunoradiometric assays (IRMA) ; competi- tive binding assays such as radioimmunoassays (RIA) ; enzyme immunoassays (EIA) ; fluorometric assays; etc. This invention will now be further illustrated by the following examples.

EXAMPLE 1 - MOLECULAR CLONING OF MOUSE BAND 3 Total poly (A ) RNA was isolated from fresh spleens of severely anemic BALB/C mice by homogeniza- tion in 5 M guanidinium isothiocyanate, purification by ultracentrifugation through a CsCl gradient, and affinity chromatography on oligo(dT) cellulose. 10 ug of this RNA was used to construct a cDNA library in the bacteriophage expression vector lambda-gtll essentially as described by Hunyh et al. Hunyh, TV. et al. , in: DNA Cloning Techniques; A Practical Approach. (D. Glover, ed., IRL Press, Oxford (1985). The endogenous Eco RI sites were not protected with Eco RI methylase. A polyclonal rabbit antibody against mouse erythrocyte Band 3 was used in conjunction with

IOC g

[ I]-protein A to screen (10 recombinant phage plaques. Fifteen antibody-positive clones were identified and subcloned into the plasmid vector, pUC13. pB33 contained an 1800 bp insert, the longest clone isolated from this library. The EcoRI RI fragment containing the entire pB33 insert was purified by agarose gel electrophoresis and labelled wit-h [alpha- 32P]dCTP by nick-translation. The labelled fragment was used to probe a ^* second mouse spleen lambda-gtll cDNA library, which was made in the following way: 10 ug of poly (A ) RNA was used to synthesize double-stranded cDNA essentially as described by Gubler. Gubler, U. and Hoffman, B.J., Gene, 25: 263-269. (1983) This cDNA was treated with EcoRI methylase and size-selected by agarose gel electrophoresis to exclude fragments less than 2.5 kb in length. Of the recombinant phage in this library, 1% contained inserts ranging from 3.5 to 4.3 kb in length which hybridized to the pB33 insert. pB399 is a pUCl3 sub-clone of the longest isolate obtained from this library. pB399 is a pSP65 subclone of the longest isolate obtained by screening the library with a 500 bp fragment EcoRI-PvuII from the 5' end of pB399.

The heavy black line in Figure 1 identifies the location of the single open reading frame encoding Band 3. Arrows denote the magnitude and direction of the fragments used to obtain the cDNA nucleotide sequence. Arrows with circles at their tails indicate that the sequence was obtained by chemical cleavage of DNA fragments end-labelled with [alpha- 3*. ~P] nucleotides at their 3' terminus (o ) or their 5' terminus (o ). Maxam, A. and Gilbert,

W. , Methods in Enzymology, 65; 499-560. (1980);

Maniatis', T. et al. , in; Molecular Cloning: A

Laboratory Manual, Cold Spring Harbor, NY (1982) . All other sequences were obtained by sub-cloning of the cDNA fragments into M13 and sequencing by the chain-termination method using [alpha- 35S]dATP.

Sanger, F. et al. , Proceedings of the National

Academy^'of Sciences, U.S.A., 74: 5463-5467 (1977) ; Biggin, M.D., Proceedings of the National Academy of Sciences, U.S.A. , 80: 3963-3965 (1983) . These fragments were obtained from restriction digests using sites originating within the cDNA (i ) or with the vector (// ) . Arrows without tails ( ) represent sequences of fragments obtained by progressive Bal 31 deletion in the case of pB33. Guo, L.- H. et al. , Nucleic Acids Research, 111: 5521-5540. (1983)

For pB399 and pB3SP4, random fragments were generated by sonication. Deninger, P.L., Analytical Biochemistry, 129; 216-223: (1983) Bal 31 diges¬ tions were performed as follows: 25 ug of plasmid was digested to completion with Nar I, which cuts pB33 at a unique site 175 bp from the 3' end of the Band 3 insert. The linearized DNA was then digested with Bal 31 (New England Biolabs) for 24 min at 30l'O using a concentration of 1 U Bal 31/ug DNA. Aliquots of the reaction mixture were removed at 3 in intervals. Following ethanol precipitation, the DNA was digested with Eco RI and ligated with Ml3 MP8 RF-DNA which had been previously digested with Eco RI and Sma I. Preparation of cDNA fragments for sonication was performed as follows: 5 ug of gel-purified restriction fragment (Eco RI/SpH I) of pB399 or (Pstl000/Pst2480) of pB3SP4 was self-ligated and sonicated as described by Deninger. The ends were repaired by treatment of the fragments with T4 polymerase in the presence of 0^'.1 mM dNTPs. The resulting DNA was fractionated on a 1.5% agarose gel. The eluted fragments were ligated with Sma I-cut M13 MP18 RF DNA. All transformations were _ done with Escherichia coli JM103. Maniatis, T. et al. , in: Molecular Cloning: A Laboratory Manual, Cold Spring Harbor. (1982)

The number of arrows in the figure represents the minimum of gels used to obtain the sequence; where there is more than one gel covering a given region of the clone, it is denoted by only a single arrow. The entire sequence was assembled with the aid of the DB programs of Staden. Staden, R. , Nucleic Acids Research, 10; 4731-4751. (1982) The database contains 174 gel readings, representing an average nucleotide redundancy of 5. The entire sequence has been obtained for both strands.

EXAMPLE 2 - ALIGNMENT OF SEQUENCES OF FRAGMENTS OF HUMAN BAND 3 PROTEIN WITH THE DEDUCED AMINO ACID SEQUENCE OF THE MOUSE PROTEIN The ALIGN program of Dayhoff et al. was used to obtain the optimal alignment using the mutation data matrix. Dayoff, M.O. et al. , Methods in Enzymology, 1: 354-545 (1983) . Shown are the alignments for an N-terminal fragment, HI, a short peptide from the N-terminus of a 14,000 dalton chymotryptic fragment, H2, and two fragments from the membrane-associated domain, H3 and H4 (see Figure 3 for their locations) . The alignment scores were compared with those obtained from 150 randomizations of the entire mouse sequence, using a break penalty of 2. The scores for the alignments were 14.6 (III), 2.66 (H2) , 15.5 (H3) and 12.2 (H4) standard deviations, respectively.

EXAMPLE 3 MOLECULAR CLONING AND ANALYSIS OF OTHER MEMBERS OF THE BAND 3 FAMILY Total poly (A ) RNA was isolated from murine mTAL-lP cells, originally derived from microdis- sected renal thick ascending limbs, in the following manner. A guanidinium thiocyanate homogenate of the cells was centrifuged through a cesium chloride cushion, (Chirgwin, J.M., Biochemistry, 1_9:5294-5299 (1979)); this was followed by affinity chroma- tography on oligo(dT) -cellulose. A cDNA library was constructed in the vector Lambda gtlO (Huynh, T.V. et al. , in DNA Cloning Techniques: A Practical Approach (D. Grover, ed., IRL Press, Oxford (1985)) by the method of Gubler and Hofman. Gubler, U. and Hof an, B.J. , Gene 25-.263-269 (1983). Recombinant phage were screened with a mix of two nick-trans¬ lated Bglll fragments of mouse erythroid Band 3, encoding bases 1748-2368 and bases 2369-3132, respectively, Kopito, R.R. and Lodish, H.F., Nature, 316:234-238 (1985). A single positive recombinant phage was purified, and its murine cDNA insert was subcloned into the plasmid pUC-13. Maniatis, T. et al. , Molecular Cloning: A Laboratory Manual, CoId Spring Harbor, NY (1982) . After plasmid amplifica¬ tion, the cDNA insert was again purified, and then subjected to "shotgun" Ml3mp8 cloning and dideoxy sequencing as described by Deininger, P.L., Analyti¬ cal Biochemistry, 129:216-223 (1983) , and by Biggin, M.D. et al. , Proceedings of the National Academy of Sciences, U.S.A., jW:3963-3965 (1983).

The full nucleotide sequence of the isolated cDNA is shown in Figure 7. The nucleotide sequence encodes 291 amino acids of the carboxy-terminal end of a'protein having approximately 66% homology with the carboxy-terminal 291 amino acid residues of mouse erythroid Band 3.

Kidney also expresses one or more mRNA trans¬ cripts which by hybridization analysis are much more highly similar to_erythroid Band 3 mRNA than that in the mouse in TAL-IP cells. Using the techniques described above, cDNA clones encoding parts of an mRNA of this type were isolated from lambda gtll cDNA libraries constructed from both mouse kidney and rat kidney mRNA.

Industrial Applicability

This invention has industrial applicability in screening for the presence or absence of genetic material (e.g., RNA or DNA) that encodes proteins which transport ions across cell membranes. In particular, it can be used to determine the presence or absence of such material encoding proteins which transport chloride ion and/or bicarbonate ion across cell membranes.

Equivalents Those skilled in the art will recognize, or be able to ascertain, using no more than routine experimentation, numerous equivalents to the specific substances and procedures described herein. Such equivalents are considered to be within the scope of this invention and are covered by the following claims.

Claims

1. cDNA encoding proteins that transport anions across cell membranes.

2. cDNA of Claim 1 encoding protein which trans- ports chloride ion across cell membranes.

3. cDNA of Claim 1 encoding protein which -trans¬ ports bicarbonate ion across cell membranes.

4. cDNA encoding murine Band 3 protein.

5. A DNA probe comprising DNA sequences homologous to DNA sequences in a gene encoding a protein that transports an anion across cell membranes.

6. A DNΛ probe of Claim 5 wherein the anion is chloride ion.

7. A DNA probe of Claim 5 wherein said protein is murine Band 3 protein.

8. A DNA probe of Claim 5 wherein said protein transports an anion across murine kidney cell membranes.

9. A recombinant vector having a DNA insert encoding protein that transports anions across cell membranes.

10. A recombinant vector containing DNA having the sequences selected from the group consisting of: a. the sequences of Figure 2; b. the sequences of Figure 7; and c. equivalent sequences.

11. Lambda-gtll vector having cDNA encoding

■ protein that transports anions across cell membranes.

12. A vector of Claim 11 in which the cDNA encodes protein that transports chloride ions across cell membranes.

13. A vector of Claim 11 in which the cDNA encodes protein that transports bicarbonate ions across cell membranes.

14. Plasmid vector having cDNA encoding murine Band 3 protein.

15. Plasmid vector having cDNA having the sequences of Figure 7 or equivalent sequences.

16. Isolated RNA encoding proteins that transport anions across cell membranes.

17. Isolated RNΛ encoding murine Band 3 protein.

18. Isolated protein encoded by the 2900 bp cDNA insert of clone pB3SP4.

19. Protein expressed by cells transformed with a recombinant vector of Claim 11.

20. Protein expressed by cells transformed with a recombinant vector of Claim 15.

21. Protein expressed by cells transformed with a recombinant vector having cDNA encoding murine Band 3 protein.

22. A protein of Claim 19 in which the recombinant vector has cDNA encoding murine Band 3 protein.,

23. An RNA probe comprising a nucleotide sequence which is essentially homologous to DNA encoding murine Band 3 protein.

24. A method of detecting a defect in chloride ion transport across cell membranes, comprising the steps of: a. obtaining cells to be assayed; b. disrupting the cells; c. isolating DNA from the disrupted cells; d. fragmenting the isolated DNA; e. separating the fragments of isolated DNA f. labeling cDNA encoding chloride ion transport protein; g. combining the fragments of "isolated

DNA of (e) and the labelled cDNA of (f); and h. detecting labelled cDNA bound to the fragments of isolated DNA.

25. A method of cloning cDNA encoding murine Band 3 protein, comprising the steps of: a. obtaining RNA from mouse anemic spleen cells; b. constructing a cDNA library by reverse transcription of the RNA from mouse anemic spleen cells; ^■ c. isolating from the cDNA library cDNA encoding murine Band 3 protein; d. ^'.inserting the isolated cDNA of (c) into an expression vector; e. transforming host cells with the ^' expression vector of (d) ; and f. culturing the transformed host cells.

26. A method of assaying mammalian cells for genetic defects associated with ion transport across mammalian cell membranes, comprising: a. obtaining DNA from the genome of cells to be assayed; b. fragmenting said DNA into a multiplicity of DNA fragments; c. contacting said DNA fragments v/ith a probe containing nucleic acid sequences homologous with a gene known to be associated with normal or defective ion transport across mammalian cell membranes under conditions which allow said probe to bind to DNA fragments having DNA sequences essentially homologous with said gene; and d. detecting any differences in DNA sequences between fragments which bind to the probe and normal mammalian cell DNA sequences coding for proteins associated with ion transport across cell membranes.

27. A method of assaying mammalian cells for genetic defects associated with ion transport across mammalian cell membranes, comprising: a. obtaining RNA from cells to be assayed; b. contacting said RNA with a probe containing nucleic acid sequences homologous with a gene known to be associated with normal or defective ion transport across mammalian cell membranes under conditions which allow said probe to bind to RNA fragments having RNA sequences essentially homologous with RNA transcribed by said gene; and c. detecting any differences between the RNA sequences which bind to the probe and normal mammalian cell RNA sequences coding for proteins associated with ion transport across cell membranes.

28. An immunochemical assay for a protein associated with transport of ions across mammalian cell membranes, comprising: a. obtaining a sample to be assayed; b. incubating said sample with antibody against said protein; and c. detecting complex formed between said antibody and said protein.

29. A cell line having the functional character- i'stics of the cell line deposited at the

American Type Culture Collection under number 53047.

30. A cell line having the functional characteris¬ tics of the cell line deposited at the American Type Culture Collection under number 67008.